NUST MISIS scientists with colleagues from BSTU. VG Shukhov and NICEM them. NF Gamalei is developing a robotic system based on machine vision that will speed up laboratory research and reduce the risk of error. This was reported to socialbites.ca at NUST MISIS.
A blood test is one of the most common types of medical research. In the preliminary stage of the material, aliquoting is done after centrifugation, that is, the liquid from the primary tube is poured into several tubes. In this case, the blood collection pipette for analysis should be immersed at different depths depending on the level of the boundaries between the fractions. Often at this stage, errors occur due to the human factor, which significantly reduces the quality of the work.
“A large number of errors that lead to a decrease in the quality of laboratory research occur precisely during the preliminary preparation of the material. A poor quality laboratory test is the need for repeated blood sampling at best, and a misdiagnosis at worst. Sergey Khalapyan, employee of the Automation and Information Control Systems Department of STI NUST MISIS, told Ru that the robotic plasma aliquoting organization will eliminate the human factor at this stage.
The developed system will automatically determine the boundary level between serum fractions in a test tube using machine vision based on a neural network. It can identify segments in the tube with 98% accuracy, then the algorithm calculates the depth to which the pipette must be dipped to collect blood serum with an error of less than 0.5 mm. This significantly improves the quality of laboratory research.
“The neural network provides high-precision image segmentation, and the algorithm developed on its basis allows you to calculate the depth at which you need to insert the pipette to get a portion. At the same time, the nature of the fractional limit is taken into account, which guarantees to obtain the maximum number of aliquots and maintain the high quality of the diagnostic work.” said Khalapyan.
In the future, the scientists will refine the system, to which they will add two more robots. The first mechanism will take a test tube containing biomaterial from one rack, transport it to the work area, wait for plasma sampling, and place the tube on another rack. The second machine will collect the plasma and distribute it into the test tubes.